This project is concerned with enhancing validity and power of neuropsychopharmacologic research and service delivery through refinements in design and analysis of controlled clinical trials, development and/or evaluation of clinical assessment and classification methodology, as well as the provision of statistical support for selected clinical investigators. Flexibility in focusing on key methodological issues as they arise has been a unique feature of this NIMH grant in the past. Evaluating neuropsychopharmacologic treatments in the broader service delivery context is a currently added emphasis that poses novel methodological problems and possibilities. Statistical issues to be examined include the best among imperfect ways of controlling for confounding factors in doctor's choice treatment settings and methodology for investigating differences in frequencies of occurrence of rare events which can only be studied in the broader service delivery context. Computer models, monte carlo simulation methods, and logical analysis will be used to examine key issues in the methodology for evaluating neuropsychopharmacologic treatments. The importance of an independent academic participation in issues that arise at the interface of industry and the FDA is stressed.